CN103867186A - Method and Apparatus for Analyzing Image Data Generated During Underground Boring or Inspection Activities - Google Patents
Method and Apparatus for Analyzing Image Data Generated During Underground Boring or Inspection Activities Download PDFInfo
- Publication number
- CN103867186A CN103867186A CN201310684922.4A CN201310684922A CN103867186A CN 103867186 A CN103867186 A CN 103867186A CN 201310684922 A CN201310684922 A CN 201310684922A CN 103867186 A CN103867186 A CN 103867186A
- Authority
- CN
- China
- Prior art keywords
- view data
- feature
- data
- defect
- probability
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000007689 inspection Methods 0.000 title claims abstract description 45
- 238000000034 method Methods 0.000 title claims abstract description 28
- 230000000694 effects Effects 0.000 title abstract 2
- 238000012549 training Methods 0.000 claims abstract description 54
- 238000010191 image analysis Methods 0.000 claims abstract description 32
- 238000004891 communication Methods 0.000 claims abstract description 19
- 230000007547 defect Effects 0.000 claims abstract description 19
- 230000006978 adaptation Effects 0.000 claims abstract description 15
- 238000004040 coloring Methods 0.000 claims description 5
- 238000012986 modification Methods 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 4
- 238000011179 visual inspection Methods 0.000 abstract description 5
- 238000005553 drilling Methods 0.000 description 17
- 238000012360 testing method Methods 0.000 description 16
- 238000005516 engineering process Methods 0.000 description 9
- 239000002689 soil Substances 0.000 description 7
- 239000010865 sewage Substances 0.000 description 6
- 238000003860 storage Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000013500 data storage Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000000007 visual effect Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000007789 gas Substances 0.000 description 3
- 238000012805 post-processing Methods 0.000 description 3
- 239000011435 rock Substances 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000004927 clay Substances 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 230000005641 tunneling Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- NRTLIYOWLVMQBO-UHFFFAOYSA-N 5-chloro-1,3-dimethyl-N-(1,1,3-trimethyl-1,3-dihydro-2-benzofuran-4-yl)pyrazole-4-carboxamide Chemical compound C=12C(C)OC(C)(C)C2=CC=CC=1NC(=O)C=1C(C)=NN(C)C=1Cl NRTLIYOWLVMQBO-UHFFFAOYSA-N 0.000 description 1
- 241000270295 Serpentes Species 0.000 description 1
- 241000656145 Thyrsites atun Species 0.000 description 1
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000013101 initial test Methods 0.000 description 1
- 239000003345 natural gas Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH DRILLING; MINING
- E21B—EARTH DRILLING, e.g. DEEP DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
- E21B47/00—Survey of boreholes or wells
- E21B47/002—Survey of boreholes or wells by visual inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
Abstract
The invention discloses a method and apparatus for analyzing image data generated during underground boring or inspection activities. The system includes a visual inspection system and an image analysis system. The visual inspection system includes an inspection camera that captures images from within an interior of at least one of a utility line and a tunnel for installing a utility line, and a first communication interface that communicates image data corresponding to the images. The image analysis system includes a second communication interface that receives the image data from the visual inspection system, a model adaptation module that modifies a classifier model based on at least one of feedback data and training data, and a classifier module that implements the classifier model to identify a plurality of features in the image data corresponding to defects and that modifies the image data according to the identified plurality of features. The defects include at least one of a cross-bore, a lateral pipe, and an imperfection.
Description
The cross reference of related application
The application require on December 17th, 2012 submit to U.S. Provisional Application the 61/738th, No. 103 and on March 13rd, 2013 submit to U.S. Provisional Application the 13/801st, the rights and interests of No. 330.Whole disclosures of above-mentioned application of quoting merge in the application by reference.
Technical field
The disclosure relates generally to subsurface utility construction field, and relates more particularly to check system and method for the view data of underground drilling operation process is analyzed.
Background technology
This part provides the background information relevant to the disclosure, and it must not be prior art.
Sometimes use any mounting technique in the multiple non-excavation mounting technique that comprises horizontal drilling technology that underground utility cause pipeline is installed.Especially, when pick is opened or excavated, have any problem in ground or when high cost, (for example for example ought there is along the path of utility line the ground obstacle of these technology of obstruction, highway, sidewalk, track or beautification of landscape) time, horizontal drilling technology provides efficient and cost-effective mode for gas, water, electricity and communication line are installed.Some horizontal drilling technology comprise underground Pneumatic drilling, auger boring, wet boring, horizontal orientation boring (HDD), impact pike, push pipe and micro-tunneling construction technology.
The process of underground Pneumatic drilling relates to starts to generate the Pneumatic drilling instrument of horizontal drilling or perforating tool to generate by the tunnel on ground along straight line path.Then, can retract utility line (for example,, for gas, water, electricity or communication) by tunnel underground to be installed on.The earth's surface obstruction that for example, can pass through existing utility line and utility line surveys and selects the path for new utility line.Opposite side at obstruction excavates two holes, and one of them hole is (entrance hole) and a hole (outlet hole) on the target endpoint in path in the starting point in path.Hole is enough large, to put into boring bar tool and operator can be worked.Hole is also enough dark, makes to keep interference-free when the ground surface that boring bar tool generates on Shi Gai tunnel, tunnel.
Boring bar tool comprises the pneumatic type boring cutter that drills soil, rock etc.Boring bar tool is connected to the compressed air of supply by flexible pipe.Steering tool and sight device are for the path along expection and towards predetermined terminal alignment boring bar tool.Then boring bar tool is activated the wall that cuts underground hole, advances and cheat by entrance, and wherein air supply hose is followed after boring bar tool.Once the progress of boring bar tool exceeds steering tool, the radio frequency receiver detecting with regard to the radio signal using being generated by the radio transmitter being built in boring bar tool is followed the trail of boring bar tool position by ground.
In the time that boring bar tool arrives target endpoint, between entrance hole and outlet hole with under the obstruction of earth's surface, generating tunnel.Remove boring bar tool and utility line is attached to air supply hose (for example,, by utility line peace is received to flexible pipe) from air supply hose.Flexible pipe and utility line by together with retract by tunnel, thereby utility line is installed to underground.
But, underground Pneumatic drilling existent defect, this shortcoming may cause being difficult to the boring for underground utility cause pipeline.For example, boring bar tool is indeflectible, once and this boring bar tool leave steering tool, operator no longer has the control of the track to boring bar tool.Therefore, for example, boring bar tool may be inclined to the path of expectation by rock and different density of soil.Even if less deflection also can cause the significant deviation with expectation path in very long distance.Therefore, boring bar tool may by mistake cross the path of other underground communal facilitys that existed.So although in fact carrying out before underground Pneumatic drilling, existing underground utility cause pipeline has been carried out to location and mark on ground, boring bar tool may pass through existing for example sanitary sewage pipeline of utility line by tunneling.Therefore, the new utility line of installing may pass existing sewage conduct.In this case, produce cross borehole, i.e. the intersection of two or more subsurface utilities.
No matter whether adopt horizontal drilling method, the significant concern point of subsurface utility construction industry be unawares tunneling by sewer line and extend thereafter utility line if natural gas line or power line are by this sewer line.Before sewer line stops up, the utility line of cross borehole may be retained in original place several months or several years.Then,, in the process of cleaning sewage conduct, utility line may be used to clear up electric power gang drill or other instruments of sewage conduct or machine cuts off, isolates or otherwise infringement.
Summary of the invention
A kind of system comprises vision inspection system and image analysis system.Described vision inspection system comprises: catch the inspection camera of image and transmit the first communication interface corresponding to the view data of described image from utility line with at least one the inside in tunnel that utility line is installed.Described image analysis system comprises: receive the second communication interface from the described view data of described vision inspection system; Revise the model adaptation module of sorter model based at least one in feedback data and training data; And classifier modules, described classifier modules realizes described sorter model to identify the multiple features corresponding to defect in described view data, and according to view data described in described multiple feature modifications of identification.Defect comprises at least one in cross borehole, horizontal pipeline and imperfection.
One method, comprising: from utility line and for install utility line tunnel at least one inside catch image.Described method also comprises with image analysis system next: receive the view data corresponding to described image; Revise sorter model based at least one in feedback data and training data; Identify the multiple features corresponding to defect in described view data with described sorter model, wherein said defect comprises at least one in cross borehole, horizontal pipeline and imperfection; And revise described view data according to identified multiple features.
Accompanying drawing explanation
The accompanying drawing of describing herein, only for the object of selected embodiment rather than all possible embodiment is described, and is not intended to limit the scope of the present disclosure.
Fig. 1 is according to the functional block diagram of the system that comprises vision inspection system and image analysis system of principle of the present disclosure;
Fig. 2 is according to the functional block diagram of the image analysis system of principle of the present disclosure; And
Fig. 3 is according to the functional block diagram of the model adaptation module of principle of the present disclosure.
The specific embodiment
Example embodiment is provided and makes the disclosure will be thoroughly, and completely scope be conveyed to those skilled in the art.Many concrete details, for example example of concrete parts, device and method is set forth to provide the understanding thoroughly to embodiment of the present disclosure.To those skilled in the art clearly, do not need to adopt concrete details, can carry out exemplifying embodiment embodiment and the two should not be interpreted as limiting the scope of the present disclosure with multiple different form.In some example embodiment, do not describe known method, known apparatus structure and known technology in detail.Now with reference to accompanying drawing, example embodiment is described more completely.
The image analysis system using together with check system is widely applicable for and in underground public facility construction industry, uses and be applicable to especially in earth drilling operating process for underground utility cause pipeline is installed.For example, check system generally includes sensor, sensor carrier and output device.Sensor is used for obtaining the inspection data relevant with the situation that operates the tunnel generating by earth drilling.In check system, can adopt the arbitrary sensor technology in multiple different sensor technology, the camera for example visual picture in tunnel being caught, and passive sensor as the touch sensing of the feature in sensing tunnel physically, can catch tunnel infrared image infrared sensor or can sensing tunnel in VOC (VOC) or the vapor sensor of the existence of other gases, or active sensor is as sonar, radar and the laser instrument that can measure the feature in tunnel.In addition, whether camera for example, catches and can identify and record another utility line through existing pipeline (the horizontal pipeline through existing pipeline being detected) for the image in the utility line to existing or pipeline (sewage conduct) after utility line can and/or be installed during current check.
Sensor carrier is applicable to combined sensor and is connected to for transporting sensor by the device in tunnel.Output device receives presenting to operator with explanation and/or otherwise recording and/or generate inspection record corresponding to the output signal of inspection data and by this output signal from sensor.In addition, output device can comprise make that operator can for example annotate user input, user interface that inspection record is added in comment etc. to.User's input can adopt any form in various ways, includes but not limited to print text, voice, timestamp and/or bookmark.In addition, output device can be configured to broadcast or announce inspection record, comprises like this operator, local government, regulator, utility company, other contractor and addressable this record of appointment recipient of the owner of title of database.In the patent cooperation treaty application of No. PCT/US2012/047290 that is to submit on July 19th, 2012, described example inspection system and method, whole content of above-mentioned application merges in the application at this.
Referring now to Fig. 1, it shows according to the disclosure and the example image analytical system 100 for using together with example vision inspection system 104.Vision inspection system 104 comprises and checks camera 108, checks that camera 108 is configured to advance before new utility line is installed by the tunnel that generates in underground Pneumatic drilling operating process and/or in the region identical with existing public utility pipeline, advances by existing public utility pipeline after new utility line is installed.In the time that camera 108 passes through tunnel, operator can observe and tunnel is carried out to the utility line of visual inspection to determine that another has existed the realtime graphic in the tunnel in display unit 112, and for example whether sanitary sewage pipeline intersects during drilling operation.Similarly, camera 108 can be through existing utility line to determine that whether this utility line is through another existing utility line.Reduce so, significantly the possibility of cross borehole and/or can cross borehole (and horizontal pipeline) detected and be proofreaied and correct.
The suitable inspection camera using together with vision inspection system of the present disclosure, can be from Ridge Tool Company of Elyria, the Ritchie Worktools Inc. in city of Illyria, OH(Ohio) obtain for example SeeSnake
(see snake
registration mark) gutter and one of Limber inspection camera and cable drum.Output from camera can comprise still frame and/or video.In addition, for obtaining from Ritchie Worktools Inc. equally the suitable display unit of observing and/or recording from the output of camera, for example SeeSnake
monitor and register.In addition, the lens of camera can change to change visual angle and/or the visual field of camera.For example, can be in conjunction with " flake " lens to catch the wall of the boring of camera periphery in the visual field of camera.In addition, check image can be recorded and/or otherwise preserve to record earth drilling operation, record and do not produce cross borehole, record and do not damage subsurface utility and/or be documented in the path in tunnel not other obstruction.
In addition, the visual inspection in the tunnel in display unit 112, by checking that the view data that camera 108 provides is transmitted to image analysis system 100.For example, check system 104 comprises communication interface 116.Only as example, communication interface operates to provide view data to image analysis system 100 according to one or more suitable wireless communication protocol, described one or more suitable wireless communication protocol includes but not limited to: wireless network (for example, Wi-Fi), mobile phone, global navigation satellites (GNSS) and/or Bluetooth protocol.Although illustrate to be independent of to check camera 108 and display unit 112, communication interface 116 can also be attached in inspection camera 108 and/or display unit 112.
Image analysis system 100 receives and can comprise the two view data of image freeze data and video data from vision inspection system 104.Only, as example, image analysis system 100 is positioned at the position away from vision inspection system 104, for example, in any suitable calculation element and/or storage device of cloud network system.But, also can in one or more parts of vision inspection system 104, implement image analysis system 100.Or, can be in vision inspection system the function of duplicating image analytical system 100.For example, can in display unit 112 and/or inspection camera 108, implement image analysis system 100.Display unit 112 can be to have for carrying out the hand-held device of user interface or the mobile device of other modes of interface with camera 108 and/or image analysis system 100.Therefore, the function of image analysis system 100 can remotely be carried out (only as example, use via the addressable server of cloud computing framework or other remote storage devices and/or treatment facility and carry out post processing) and/or can on job site, carry out by being configured to implement the local device of image analysis system 100 (for example, post processing and/or in real time).
Image analysis system 100 carries out the view data part of graphical analysis with the identification any cross borehole of indication and/or horizontal pipeline to view data.For example, image analysis system 100 implementation models, this model is classified to multiple features of the indication cross borehole in image data frame and/or horizontal pipeline and is the each characteristic allocation probability in existence or the multiple features that do not have cross borehole and/or horizontal pipeline for image data frame.Graphical analysis can also be identified utility line or carrier and/or for the imperfection of the tunnel of utility line or carrier or the other types of boring is installed.Imperfection can include, but are not limited to the inconsistency of tunnel surface.Only, as example, inconsistency may be for example, for example, to be caused by the broken and/or straight surface crossing with circular surfaces (periphery that, pipeline or other straight objects pass through tunnel) of soil and/or pipeline in the space in surrounding soil, tunnel (, vitrified-clay pipe).
In addition, image analysis system 100 can carry out graphical analysis to detect cross borehole, horizontal pipeline, imperfection etc. no matter vision inspection system 104 whether for this object.For example, vision inspection system 104 can be by directly observing or processing in real time or post processing by use image analysis system 100 in display unit 112, for example, for identifying and locate other features of utility line (, leader pipe, gutter etc.).But, attempting identification when other features, the cross borehole of image analysis system 100 in still can recognition image data, horizontal pipeline, imperfection etc.
Referring now to Fig. 2, it shows example image analytical system 200.Image analysis system 200 is for example communicated by letter with the vision inspection system 104 of Fig. 1 via communication interface 204.For example, or as mentioned above, image analysis system 200 can carry out integrated (, carrying out integrated with display unit 112, inspection camera 108 and/or another device of vision inspection system 104) with vision inspection system 104.For example, communication interface 204 receives the view data from vision inspection system 104.View data is stored in image data memory 208.Only, as example, image data memory 208 comprises the nonvolatile memory of storing image data.View data comprises video data and/or Still image data.
Each frame is classified corresponding to the feature of cross borehole and/or horizontal pipeline and according to the feature of identification in each frame of classifier modules 212 recognition image data.For example, classifier modules 212 realizes sorter model, and this sorter model is analyzed and classifies each frame according to the feature in frame.Only, as example, each frame is assigned with one or more mark, and described mark includes but not limited to " horizontal pipeline ", " without horizontal pipeline ", " cross borehole " and/or " without cross borehole ".Classifier modules 212 is stored in the view data of classification in the image data memory 216 of classification.The view data of classification is offered vision inspection system 200 or another device or user's (for example,, according to request) by communication interface 204.
Relatively, training data can comprise and have feature the training image data (for example, training video) of the various combinations of (for example, cross borehole, without cross borehole, horizontal pipeline and/or without horizontal pipeline).Model adaptation module 220 is from training image extracting data feature and correspondingly each frame is marked the training data of (for example, using a model) and storage classification.The training data of classification and the test data of the actual characteristic of indication training image data are compared the result with assessment models by model adaptation module 220.Model adaptation module 220 is upgraded the model being used by classifier modules 212 according to this result.
Referring now to Fig. 3, example model adaptation module 300 comprises training image data sorter module 304, training and testing data storage module 308 and outcome evaluation module 312.Training image data sorter module receives training image data and feedback data, extracts the feature of indication cross borehole and/or horizontal pipeline and the training image data of classification are offered to training and testing data storage module 308 from training image data.Only, as example, the training image data of classification can be divided into two groups, and wherein first group of training image data comprises corresponding to feature and second group of training image data of cross borehole and comprise the feature corresponding to horizontal pipeline.Outcome evaluation module 312 is by the training image data of classification and test data compares and performance based on this comparative assessment model.The output (for example, model is adjusted signal) of outcome evaluation module 312 is indicated the performance of model and is provided for classifier modules 212 correspondingly this model is adjusted.
Training image data can comprise the multiple videos that are for example arranged in the different sets that comprises different individual features.For example, training image data can comprise horizontal inspection training set, and this laterally checks that training set comprises more than first video with horizontal pipeline and more than second video without horizontal pipeline.Relatively, training image data can also comprise that cross borehole checks training set, and this cross borehole checks that training set comprises more than first video with cross borehole and more than second video without cross borehole.
Training image and data sorter module 304 are for each in horizontal pipeline and cross borehole, and extraction can be indicated one or more feature of horizontal pipeline or cross borehole.Feature can include but not limited to the discrete histogram of parallel lines, colouring information, K mean cluster and/or gradient magnitude.For example, the parallel lines in image data frame can be indicated horizontal pipeline.Classifier modules 304 can realize edge detector with the Hough line in inspection image data, select for example, for example, a strongest Hough line in a Hough line in Hough line L1 (in Hough line the strongest a Hough line) and selection (, in the threshold value of L1 for example in 5 °) Hough line L2 parallel with L1.May cause under the situation for uneven line, can applying perspective and analyzing (Perspective analysis) in actual view data along the direction trend of camera at pipeline.Can use Kalman filter to follow the tracks of adjusts corresponding to the probability of horizontal pipeline line L1 and line L2.For example, Kalman filter is followed the tracks of the end position that can follow the tracks of to predict from initial test point pipeline to doubtful horizontal pipeline.If the end position of prediction is corresponding to the actual feature detecting in frame, this frame may comprise horizontal pipeline.
Colouring information can be indicated horizontal pipeline and/or cross borehole.For example, training image and data sorter module 304 for example can realize HSV histogram, to be identified in the amount color of the known color of the public utility pipeline of certain type (, corresponding to) of selected color in the part of image data frame.
K mean cluster can be indicated horizontal pipeline and/or cross borehole.For example, on histogram single Gaussian Profile of (for example,, on tone) can distribute corresponding to for example, double gauss without (, on tone) on cross borehole and histogram can be corresponding to cross borehole.
The discrete histogram of gradient magnitude can for rim detection and indication horizontal pipeline and cross borehole the two.Only, as example, can for example, after application Gaussian Blur and edge detector (, Canny edge detector) removal noise, calculate discrete histogram.In this histogram, the relatively strong edge of peak value indication more than threshold value.
Extracting after feature from training image Frame, training image and data sorter module 304 mark (classifying via operator/user input) to the each frame in frame.Mark can comprise " horizontal pipeline ", " without horizontal pipeline ", " cross borehole " and/or " without cross borehole " and can comprise sub-mark, for example " approach horizontal pipeline ", the type (for example, sandy soil, clay, rock soil etc.) of " approaching cross borehole " and soil.For model adaptation module 300, mark is manually applied (, by human operator/user).In other words, operator observes each frame and the visual properties based in image marks frame.
For the each feature in the feature of extracting, the mark of model adaptation module 300 based on being distributed by operator carrys out assigned characteristics corresponding to cross borehole, without cross borehole, horizontal pipeline and/or the probability without horizontal pipeline.For example, feature (for example, two the strong parallel lines) how many times of classifier modules 304 based on corresponding extraction is finally labeled as by operator the probability that horizontal pipeline distributes two strong parallel lines indication horizontal pipelines.Relatively, feature (for example, two the strong parallel lines) how many times of classifier modules 304 based on corresponding extraction is finally labeled as without horizontal pipeline and distributes two strong parallel lines not correspond to the probability of horizontal pipeline (, without horizontal pipeline) by operator.Therefore, the each feature in the feature of extraction is assigned with corresponding probability.Then, according to the training image data of mark storage classification.For example, can pass through frame number (for example, 1,2 ... n) with training and testing data storage 308 in corresponding mark and indicate each feature, corresponding to the data of the probability of corresponding mark, each frame is carried out to index.
Test data is stored in training and testing data storage 308 together with training image data.Test data frame corresponding to test data is fewer than training image data.In other words the view data that, is stored in the major part in training and testing data storage 308 is corresponding to training image data rather than test data.As mentioned above, by operator, training image data are marked.By contrast, operator does not mark test data.But, according to model, the test data that may comprise the image data frame identical with the part of training image data is analyzed.The training image data of the analysis result (, the mark of distribution and/or probability) of test data and mark compare to determine the accuracy of model.Can adjust model based on this accuracy.For example, outcome evaluation module 312 can be determined the error rate being associated with the assessment of test data and/or the vision response test being associated with multiple assessments.
Referring again to Fig. 2, classifier modules 212 is analyzed (, from image, extracting feature) and the feature based on extracting and training image data to the view data receiving from image data memory 208 and is calculated each frame and comprise cross borehole, do not comprise cross borehole, comprise horizontal pipeline or do not comprise the probability of horizontal pipeline.For example, frame comprises that the probability of cross borehole can comprise the combination corresponding to the each probability in the probability of cross borehole of each feature in the feature detecting in frame.Frame does not comprise that the probability of cross borehole can comprise the combination corresponding to the each probability in the probability without cross borehole of each feature in the feature detecting in frame.Frame comprises that the probability of horizontal pipeline can comprise the combination corresponding to the each probability in the probability of horizontal pipeline of each feature in the feature detecting in frame.Frame does not comprise that the probability of horizontal pipeline can comprise the combination corresponding to the each probability in the probability without horizontal pipeline of each feature in the feature detecting in frame.
Only as example, can be according to the whole bag of tricks, for example Naive Bayes Classification is calculated probability.For example, the probability that Naive Bayes Classification can be based on distributing to parallel lines detection, colouring information and discrete histogrammic probability calculation frame and comprise horizontal pipeline.If calculate probability be greater than threshold value, classifier modules 212 by horizontal pipeline subcarrier label assignments to frame.Only, as example, threshold value can be (for example, 50%) of fixing and/or can be adjusted into comparatively responsive (, reducing) or more insensitive (, improving).Classifier modules 212 determine whether in a similar fashion by without horizontal pipeline, cross borehole, without the subcarrier label assignments of cross borehole to frame.In other embodiment, probability may simply be individual features probability and, average or any other combination.
In addition,, for some frames, for example, may all be less than corresponding threshold value for all probability of distribute labels (, horizontal pipeline, without horizontal pipeline, cross borehole and without cross borehole).Therefore, frame may not have the qualification of the arbitrary mark in mark.For this frame, classifier modules 212 can be carried out " arest neighbors " and be calculated to distribute one or more mark.For example, classifier modules 212 can the feature based on extracting determine which training image Frame is the most similar to this frame.The mark of classifier modules 212 based on distributing to immediate training image Frame marks this frame.
Aforesaid explanation is only illustrative in itself and is never intended to limit the disclosure and application or purposes.Can implement in a variety of forms teaching widely of the present disclosure.Therefore,, although the disclosure comprises specific example, true scope of the present disclosure should not be so limited because by research accompanying drawing, manual and claims, it is obvious that other modification will become.For the sake of clarity, use in the accompanying drawings the similar element of identical designated.As used in this manner, at least one in phrase A, B and C is appreciated that the logic (A or B or C) that means to use nonexcludability logic OR.Should be understood that, in the situation that not changing principle of the present disclosure, can carry out one or more step in manner of execution according to different order (or concurrently).
Ground as used herein, term module can refer to a following part in every or comprise following every: special IC (ASIC); Discrete circuit; Integrated circuit; Combinational logic circuit; Field programmable gate array (FPGA); (that share, special or group) processor of run time version; Other suitable hardware componenies of described function are provided; Or the some or all combination in above-mentioned, for example SOC(system on a chip).Term module can comprise (that share, special or group) memory, and this memory stores is by the performed code of processor.
As above-mentioned use, term code can comprise software, firmware and/or microcode and can refer to program, routine, function, class and/or object.As above make land used, term is shared some codes or the whole code that refer to from multiple modules and can be used single (sharing) processor to carry out.In addition can store by single (sharing) memory from some or all codes of multiple modules.As above make land used, term group refers to from some codes of individual module or whole code and can carry out with one group of processor.In addition, can use storage stack to store from some codes or whole code of individual module.
Equipment as herein described and method can be partially or fully by being implemented by one or more performed computer program of one or more processor.Computer program comprises processor executable, and this processor executable is stored on the tangible computer-readable medium of at least one non-transient state.Computer program can also comprise and/or rely on the data of storage.The non-limiting example of the tangible computer-readable medium of non-transient state comprises nonvolatile memory, volatile memory, magnetic memory and optical memory.
Claims (20)
1. a system, comprising:
Vision inspection system, described vision inspection system comprises:
Check camera, described inspection camera from utility line and for install utility line tunnel at least one inside catch image, and
The first communication interface, described the first communication interface transmits the view data corresponding to described image; And
Image analysis system, described image analysis system comprises:
Second communication interface, described second communication interface receives the described view data from described vision inspection system,
Model adaptation module, described model adaptation module is revised sorter model based at least one of feedback data and training data, and
Classifier modules, described classifier modules realizes described sorter model to identify the multiple features corresponding to defect in described view data, wherein said defect comprises at least one in cross borehole, horizontal pipeline and imperfection, and described classifier modules is revised described view data according to identified multiple features.
2. system according to claim 1, wherein, described image analysis system is be positioned at away from the position of described vision inspection system and be integrated at least one among described vision inspection system.
3. system according to claim 1, wherein, described image analysis system receives the described view data from described vision inspection system via cloud network system.
4. system according to claim 1, wherein, described sorter model is classified to the each feature in described multiple features, and each feature in wherein said multiple feature is indicated at least one defect in described defect.
5. system according to claim 4, wherein, described sorter model is the each characteristic allocation probability in described multiple feature, and each probability in wherein said probability is corresponding to the probability that has the corresponding a kind of defect in described defect in described view data.
6. system according to claim 1, wherein, comprises the frame distribute labels of the described view data for comprising a kind of feature in identified feature according to view data described in identified feature modification.
7. system according to claim 6, wherein, described mark is included in horizontal pipeline in described frame, without horizontal pipeline, cross borehole and without at least one the indication in cross borehole.
8. system according to claim 1, wherein, described training data comprises the training image data that include described multiple features.
9. system according to claim 8, wherein, the described multiple features whether described model adaptation module identifies in described training image data based on described sorter model are revised described sorter model.
10. system according to claim 1, wherein, described multiple features comprise at least one in parallel lines, colouring information, K mean cluster and gradient.
11. 1 kinds of methods, comprising:
From utility line and for install utility line tunnel at least one inside catch image; And
Next with image analysis system:
Receive the view data corresponding to described image;
Revise sorter model based at least one in feedback data and training data;
Identify the multiple features corresponding to defect in described view data with described sorter model, wherein said defect comprises at least one in cross borehole, horizontal pipeline and imperfection; And
Revise described view data according to identified multiple features.
12. methods according to claim 11, wherein, described image analysis system is be positioned at away from the position of described vision inspection system and be integrated at least one among described vision inspection system.
13. methods according to claim 11, wherein, described image analysis system receives described view data via cloud network system.
14. methods according to claim 11, also comprise and use described sorter model to classify to the each feature in described multiple features, the each feature in wherein said multiple features is indicated at least one defect in described defect.
15. methods according to claim 14, also comprise that using described sorter model is the each characteristic allocation probability in described multiple feature, the each probability in wherein said probability is corresponding to the probability that has the corresponding a kind of defect in described defect in described view data.
16. methods according to claim 11, wherein, comprise the frame distribute labels of the described view data for comprising a kind of feature in identified feature according to view data described in identified feature modification.
17. methods according to claim 16, wherein, described mark comprises horizontal pipeline in described frame, without horizontal pipeline, cross borehole and without at least one the indication in cross borehole.
18. methods according to claim 11, wherein, described training data comprises the training image data that include described multiple features.
19. methods according to claim 18, also comprise that the described multiple features that whether identify in described training image data based on described sorter model revise described sorter model.
20. methods according to claim 11, wherein, described multiple features comprise at least one in parallel lines, colouring information, K mean cluster and gradient.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US201261738103P | 2012-12-17 | 2012-12-17 | |
| US61/738,103 | 2012-12-17 | ||
| US13/801,330 US9219886B2 (en) | 2012-12-17 | 2013-03-13 | Method and apparatus for analyzing image data generated during underground boring or inspection activities |
| US13/801,330 | 2013-03-13 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103867186A true CN103867186A (en) | 2014-06-18 |
| CN103867186B CN103867186B (en) | 2018-05-25 |
Family
ID=49766866
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310684922.4A Expired - Fee Related CN103867186B (en) | 2012-12-17 | 2013-12-13 | The method and apparatus of the image data generated in analysis earth drilling or Survey Operations |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US9219886B2 (en) |
| EP (1) | EP2743447A1 (en) |
| JP (1) | JP6005026B2 (en) |
| CN (1) | CN103867186B (en) |
| CA (1) | CA2834268A1 (en) |
| RU (1) | RU2013155978A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108241870A (en) * | 2016-12-23 | 2018-07-03 | 赫克斯冈技术中心 | For distributing specific class method for distinguishing interested in measurement data |
| CN109145951A (en) * | 2018-07-23 | 2019-01-04 | 昆明理工大学 | A kind of tunnel lining structure duty status evaluation method based on Bayesian network |
| CN109462999A (en) * | 2016-05-23 | 2019-03-12 | Ani有限公司 | Vision inspection method by data balancing based on study and the vision inspection device using it |
| CN111670357A (en) * | 2018-01-31 | 2020-09-15 | 三星电子株式会社 | Visual inspection management method and visual inspection system |
| CN112840383A (en) * | 2018-10-24 | 2021-05-25 | 富士通先端科技株式会社 | Banknote inspection device, banknote inspection method, and banknote inspection program |
| CN113256551A (en) * | 2021-01-21 | 2021-08-13 | 中国煤炭科工集团太原研究院有限公司 | Roadway roof rigid belt drilling identification and positioning system and method based on machine vision |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10427290B2 (en) | 2017-07-18 | 2019-10-01 | General Electric Company | Crawler robot for in situ gap inspection |
| US20140292616A1 (en) * | 2013-03-28 | 2014-10-02 | Nvidia Corporation | Computer monitor equalization using handheld device |
| JP6428089B2 (en) | 2014-09-24 | 2018-11-28 | 日亜化学工業株式会社 | Light emitting device |
| US10469511B2 (en) * | 2016-03-28 | 2019-11-05 | Cisco Technology, Inc. | User assistance coordination in anomaly detection |
| CN106157182B (en) * | 2016-07-07 | 2021-03-23 | 云南省交通规划设计研究院有限公司 | Method for judging quality defect degree of inverted arch of highway tunnel |
| US10596713B2 (en) | 2017-07-18 | 2020-03-24 | General Electric Company | Actuated sensor module and method for in situ gap inspection robots |
| US10434641B2 (en) | 2017-07-18 | 2019-10-08 | General Electric Company | In situ gap inspection robot system and method |
| US10603802B2 (en) | 2017-07-18 | 2020-03-31 | General Electric Company | End region inspection module and method for in situ gap inspection robot system |
| US10427734B2 (en) | 2017-07-18 | 2019-10-01 | General Electric Company | Omnidirectional traction module for a robot |
| US20200005406A1 (en) * | 2018-06-27 | 2020-01-02 | Hydromax USA, LLC | Cross-bore risk assessment and risk management tool |
| CN109583486A (en) * | 2018-11-21 | 2019-04-05 | 银河水滴科技（北京）有限公司 | A kind of method and device detecting environmental abnormality region to be measured |
| US11592594B2 (en) * | 2020-04-13 | 2023-02-28 | X Development Llc | Subsurface lithological model with machine learning |
| BR102020012943A2 (en) * | 2020-06-24 | 2022-01-04 | Petróleo Brasileiro S.A. - Petrobras | COMPUTATIONAL SCRIPT FOR IMAGE TREATMENT AND ITS APPLICATION IN EASY IMAGE DETERMINATION METHOD |
| FR3122924A1 (en) * | 2021-05-12 | 2022-11-18 | Veolia Environnement | Fault identification system in a pipeline |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030023404A1 (en) * | 2000-11-22 | 2003-01-30 | Osama Moselhi | Method and apparatus for the automated detection and classification of defects in sewer pipes |
| US20030118230A1 (en) * | 2001-12-22 | 2003-06-26 | Haoshi Song | Coiled tubing inspection system using image pattern recognition |
| US20080210024A1 (en) * | 2007-02-28 | 2008-09-04 | Merlo Stephen A | Remote pipe inspection |
| EP2192538A2 (en) * | 2008-11-21 | 2010-06-02 | Emerson Electric Co. | System for sharing video captured at jobsite |
Family Cites Families (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS61193090A (en) * | 1985-02-22 | 1986-08-27 | Nippon Kokan Kk <Nkk> | Detector for underground obstacle |
| US4974168A (en) | 1988-04-19 | 1990-11-27 | Cherne Industries, Inc. | Automatic pipeline data collection and display system |
| US5331550A (en) | 1991-03-05 | 1994-07-19 | E. I. Du Pont De Nemours And Company | Application of neural networks as an aid in medical diagnosis and general anomaly detection |
| DE4328031A1 (en) | 1992-09-24 | 1994-03-31 | Deutsche Aerospace | Fault detection system for drainage pipes - using pig supporting video camera for visual examination and radar antenna |
| US5454276A (en) | 1993-07-30 | 1995-10-03 | Wernicke; Timothy K. | Multi-directional magnetic flux pipe inspection apparatus and method |
| FR2712663B1 (en) | 1993-11-18 | 1996-01-19 | Val De Marne General Conseil | Apparatus for inspecting the physical state of pipes that are not accessible or can be visited by humans. |
| JP2741653B2 (en) * | 1994-06-03 | 1998-04-22 | トシマ建設工業株式会社 | Drilling tube for chemical injection |
| US5742517A (en) | 1995-08-29 | 1998-04-21 | Integrated Computer Utilities, Llc | Method for randomly accessing stored video and a field inspection system employing the same |
| US6175380B1 (en) | 1996-08-28 | 2001-01-16 | Peninsular Technologies, Llc | Method for randomly accessing stored imagery and a field inspection system employing the same |
| US5892163A (en) | 1998-01-27 | 1999-04-06 | Geophysical Survey Systems, Inc. | Sewer and pipeline inspection transporter tool |
| US6239593B1 (en) | 1998-09-21 | 2001-05-29 | Southwest Research Institute | Method and system for detecting and characterizing mechanical damage in pipelines using nonlinear harmonics techniques |
| US6650779B2 (en) | 1999-03-26 | 2003-11-18 | Georgia Tech Research Corp. | Method and apparatus for analyzing an image to detect and identify patterns |
| US7131344B2 (en) | 2002-03-13 | 2006-11-07 | Burn-Am Co., Ltd. | Device and method for inspecting inside of underground pipe line and method of inspecting concrete on inside of underground pipe line for deterioration |
| US7181985B2 (en) | 2005-05-27 | 2007-02-27 | Breval Technical Services Limited | Conduit inspection apparatus and method |
| WO2011080590A2 (en) | 2009-12-30 | 2011-07-07 | Sewervue Technology Corp. | Apparatus and method for inspection of underground sewer pipes |
| US8988969B2 (en) | 2010-04-23 | 2015-03-24 | Underground Imaging Technologies, Inc. | Detection of cross bores involving buried utilities |
| US9399910B2 (en) | 2011-07-21 | 2016-07-26 | Emerson Electric Co. | Inspection system and method for use in underground boring operations |
-
2013
- 2013-03-13 US US13/801,330 patent/US9219886B2/en not_active Expired - Fee Related
- 2013-11-22 JP JP2013241605A patent/JP6005026B2/en not_active Expired - Fee Related
- 2013-11-27 CA CA2834268A patent/CA2834268A1/en not_active Abandoned
- 2013-12-03 EP EP13195562.7A patent/EP2743447A1/en not_active Withdrawn
- 2013-12-13 CN CN201310684922.4A patent/CN103867186B/en not_active Expired - Fee Related
- 2013-12-17 RU RU2013155978/08A patent/RU2013155978A/en not_active Application Discontinuation
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030023404A1 (en) * | 2000-11-22 | 2003-01-30 | Osama Moselhi | Method and apparatus for the automated detection and classification of defects in sewer pipes |
| US20030118230A1 (en) * | 2001-12-22 | 2003-06-26 | Haoshi Song | Coiled tubing inspection system using image pattern recognition |
| US20080210024A1 (en) * | 2007-02-28 | 2008-09-04 | Merlo Stephen A | Remote pipe inspection |
| EP2192538A2 (en) * | 2008-11-21 | 2010-06-02 | Emerson Electric Co. | System for sharing video captured at jobsite |
Non-Patent Citations (1)
| Title |
|---|
| SUNIL K SINHA: "Automated underground pipe inspection using a unified image processing and artificial intelligence methodology", 《HTTP://HDL.HANDLE.NET/10012/578》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109462999A (en) * | 2016-05-23 | 2019-03-12 | Ani有限公司 | Vision inspection method by data balancing based on study and the vision inspection device using it |
| CN109462999B (en) * | 2016-05-23 | 2021-05-07 | Ani有限公司 | Visual inspection method based on learning through data balance and visual inspection device using same |
| CN108241870A (en) * | 2016-12-23 | 2018-07-03 | 赫克斯冈技术中心 | For distributing specific class method for distinguishing interested in measurement data |
| CN108241870B (en) * | 2016-12-23 | 2022-01-25 | 赫克斯冈技术中心 | Method for assigning a specific category of interest within measurement data |
| CN111670357A (en) * | 2018-01-31 | 2020-09-15 | 三星电子株式会社 | Visual inspection management method and visual inspection system |
| CN109145951A (en) * | 2018-07-23 | 2019-01-04 | 昆明理工大学 | A kind of tunnel lining structure duty status evaluation method based on Bayesian network |
| CN109145951B (en) * | 2018-07-23 | 2022-04-12 | 昆明理工大学 | Bayesian network-based tunnel lining structure service condition evaluation method |
| CN112840383A (en) * | 2018-10-24 | 2021-05-25 | 富士通先端科技株式会社 | Banknote inspection device, banknote inspection method, and banknote inspection program |
| CN112840383B (en) * | 2018-10-24 | 2024-03-08 | 富士通先端科技株式会社 | Banknote checking device, banknote checking method, and banknote checking program |
| CN113256551A (en) * | 2021-01-21 | 2021-08-13 | 中国煤炭科工集团太原研究院有限公司 | Roadway roof rigid belt drilling identification and positioning system and method based on machine vision |
| CN113256551B (en) * | 2021-01-21 | 2023-03-14 | 中国煤炭科工集团太原研究院有限公司 | Roadway roof rigid belt drilling identification and positioning system and method based on machine vision |
Also Published As
| Publication number | Publication date |
|---|---|
| JP6005026B2 (en) | 2016-10-12 |
| US9219886B2 (en) | 2015-12-22 |
| RU2013155978A (en) | 2015-06-27 |
| CA2834268A1 (en) | 2014-06-17 |
| US20140168408A1 (en) | 2014-06-19 |
| JP2014122887A (en) | 2014-07-03 |
| EP2743447A1 (en) | 2014-06-18 |
| CN103867186B (en) | 2018-05-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103867186A (en) | Method and Apparatus for Analyzing Image Data Generated During Underground Boring or Inspection Activities | |
| Weidner et al. | Classification methods for point clouds in rock slope monitoring: A novel machine learning approach and comparative analysis | |
| CN107655457B (en) | A kind of Geological Hazards of debris recognition methods based on remote sensing satellite image | |
| US20190087939A1 (en) | Inferring petrophysical properties of hydrocarbon reservoirs using a neural network | |
| Lato et al. | Comparison of airborne laser scanning, terrestrial laser scanning, and terrestrial photogrammetry for mapping differential slope change in mountainous terrain | |
| KR102399227B1 (en) | Managing system of underground facilities using gpr detector | |
| KR101895835B1 (en) | Ground penetrating radar survey system | |
| US20150218936A1 (en) | Down-hole monitoring and survey system | |
| CN104239921A (en) | Underground pipeline marking and real-time depth measurement system | |
| US20220067420A1 (en) | System and Method for Detection of Carbonate Core Features from Core Images | |
| JP6083965B2 (en) | Buried object exploration system | |
| CN107843208A (en) | Mine roadway contour sensing method and system | |
| KR102074462B1 (en) | System for detecting sinkhole using infrared iamge analysis and method for detecting sink hole using it | |
| KR20220029396A (en) | Automation device and method for evaluating Rock Mass Rating(RMR) using deep learning models | |
| CN103942576A (en) | Method for identifying coal and rock through airspace multiscale random characteristics | |
| US20130085676A1 (en) | Processing of Geological Data | |
| US20230038924A1 (en) | Core model augmented reality | |
| KR102205080B1 (en) | Method for process of gpr data | |
| JP2018091003A (en) | Underground survey system | |
| CN208477368U (en) | A kind of contactless asphalt mixing plant monitoring management system | |
| Govada et al. | Road deformation detection | |
| KR102446320B1 (en) | Apparatus for Map-based Object Tracking and Event Detection Based on a plurality of Cameras and Driving Method Thereof | |
| Pradhananga et al. | Spatio-temporal safety analysis of construction site operations using GPS data | |
| US11802474B2 (en) | Formation-cutting analysis system for detecting downhole problems during a drilling operation | |
| JP7145249B2 (en) | Buried object discrimination system and buried object discrimination method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180525 Termination date: 20211213 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |